The Dual-Specificity Tyrosine-Regulated Kinase 1B (DYRK1B, also referred to as Minibrain-Regulated Kinase MIRK) belongs to the DYRK family of serine/threonine kinases, which, based on sequence and structural homologies, can be divided into three subgroups: the YAK group with no members in the animal kingdom, the DYRK1 and the DYRK2 subgroup. The two DYRK1 subgroup members DYRK1A and DYRK1B share 85% identity at the amino acid level, though expression and functional characteristics are distinct (Aranda et al., FASEB J. 2011 February; 25(2):449-462.).
The human DYRK1B gene encodes a 69 kDa protein with 629 amino acids in length. Alternative splicing and differential promoter engagement can yield two additional, slightly shorter DYRK1B isoforms with differential expression patterns, the shorter of which lacking kinase activity (Leder et al., Biochem J. 2003 Jun. 15; 372(Pt 3):881-888.).
The regulation of DYRK1B catalytic activity and function is not entirely understood. Given the extensive sequence similarity to DYRK1A, the intrinsic regulatory and catalytic properties of DYRK1B can—to a certain extent—be inferred from studies of DYRK1A regulation. DYRK family members are arginine-directed serine/threonine kinases, with DYRK1B phosphorylating either serine in the consensus substrate sequence SPSxxR (Friedman, J Cell Biochem. 2007 Oct. 1; 102(2):274-279; Himpel et al., J Biol Chem. 2000 Jan. 28; 275(4):2431-2438.). Activation of DYRK1B involves an intramolecular tyrosine (Y) auto-phosphorylation of the second tyrosine of an Y×Y motif in the conserved kinase domain and activation loop, respectively (Becker et al., FEBS J. 2011 January; 278(2):246-256; Himpel et al., J Biol Chem. 2000 Jan. 28; 275(4):2431-2438; Nolen at al., Mol Cell. 2004 Sep. 10; 15(5):661-675.). Notably, this Y-phosphorylation-dependent activation step occurs only during translation of DYRK kinases, resulting in activated DYRK proteins with serine/threonine kinase activity (Lochhead et al., Cell. 2005 Jun. 17; 121(6):925-936.). This implies that additional processes including protein-protein interactions, further post-translational modifications and/or subcellular localization control the response of DYRK kinases to extracellular signals. For instance, signaling via RAS/RAC/MKK3 is able to stimulate DYRK1B activity in certain cellular contexts such as in pancreatic cancer cells and interaction of DYRK1A with 14-3-3 proteins significantly enhances DYRK kinase activity (Deng et al., J Biol Chem. 2003 Oct. 17; 278(42):41347-41354; Jin et al., J Biol Chem. 2005 Dec. 23; 280(51):42097-42105; Jin et al., Cancer Res. 2007 Aug. 1; 67(15):7247-7255; Kim et al., Biochem Biophys Res Commun. 2004 Oct. 15; 323(2):499-504.).
During normal development DYRK1B expression is preferentially restricted to testes and muscle cells (Leder et al., Biochem J. 2003 Jun. 15; 372(Pt 3):881-888; Leder et al., Biochem Biophys Res Commun. 1999 Jan. 19; 254(2):474-479.). In muscle its expression is regulated by RHO and basic Helix-Loop-Helix transcription factors via binding to an E-box in the DYRK1B promoter (Deng et al., J Biol Chem. 2003 Oct. 17; 278(42):41347-41354; Friedman, J Cell Biochem. 2007 Oct. 1; 102(2):274-279.). Inhibition and overexpression studies suggest a pro-differentiation activity of DYRK1B in myoblast differentiation by enhancing the expression of myogenic transcription factors such as Mef2 (Deng et al., J Biol Chem. 2005 Feb. 11; 280(6):4894-4905.). This effect is mediated by DYRK1B-dependent phosphorylation of class II histone deacetylases (HDAC) thereby freeing Mef2 from complexes with inhibitory HDACs and allowing Mef2 to exert its pro-myogenic function (Deng et al., J Biol Chem. 2005 Feb. 11; 280(6):4894-4905.). In addition to its role in myoblast differentiation, DYRK1B also controls cell cycle arrest by phosphorylation-dependent destabilization of D-type cyclins and stabilization of cell cycle inhibitors including p27 and p21 (Deng et al., J Biol Chem. 2004 May 21; 279(21):22498-22504; Ewton et al., Mol Cancer Ther. 2011 November; 10(11):2104-2114; Mercer et al., J Biol Chem. 2005 Jul. 8; 280(27):25788-25801; Zou et al., J Biol Chem. 2004 Jun. 25; 279(26):27790-27798.). Like DYRK1A, for which a larger number of substrate proteins has already been identified, DYRK1B can act as co-activator of the FOXO1a transcription factor, thereby regulating glucose-6-phosphatase expression (von Groote-Bidlingmaier et al., Biochem Biophys Res Commun. 2003 Jan. 17; 300(3):764-769.).
DYRK1B deficient mice do not display any evident developmental defects and survive several weeks post birth (Leder et al., Biochem J. 2003 Jun. 15; 372(Pt 3):881-888.). Details of the DYRK1B mutant phenotype remain to be reported. By contrast, DYRK1A deficiency in mice results in an embryonic lethal phenotype (Fotaki et al., Mol Cell Biol. 2002 September; 22(18):6636-6647.).
DYRK1B in Cancer
Several recent studies have implicated DYRK1B as a putative oncogenic factor in different cancer entities. DYRK1B localizes to the chromosomal region 19q13, which is frequently amplified in pancreatic and ovarian cancers (Friedman, J Cell Biochem. 2007 Oct. 1; 102(2):274-279; Karhu et al., Genes Chromosomes Cancer. 2006 August; 45(8):721-730; Lee et al., Cancer Res. 2000 Jul. 1; 60(13):3631-3637.). Accordingly, DYRK1B is strongly expressed in a fraction of pancreatic and ovarian cancer cell lines (Friedman, J Cell Biochem. 2007 Oct. 1; 102(2):274-279; Hu and Friedman, Genes Cancer. 2010 Aug. 1; 1(8):803-811.). Notably, in pancreatic cancer DYRK1B acts as survival effector kinase downstream of RAS-RAC1 to promote viability and clonal growth of cancer cells (Jin et al., Cancer Res. 2007 Aug. 1; 67(15):7247-7255.). In addition, several in vitro studies suggest that DYRK1B has pro-oncogenic function in colon cancer, osteosarcoma and rhabdomyosarcoma (RMS). RNA interference and overexpression studies demonstrated a pro-survival role of DYRK1B in colon cancer, osteosarcoma, RMS and pancreatic cancer (Deng et al., Cancer Res. 2009 Apr. 15; 69(8):3317-3324; Deng et al., Cancer Res. 2006 Apr. 15; 66(8):4149-4158; Friedman, Sarcoma. 2011; 2011:260757, doi:10.1155/2011/260757; Jin et al., Cancer Res. 2007 Aug. 1; 67(15):7247-7255; Mercer et al., Cancer Res. 2006 May 15; 66(10):5143-5150; Yang et al., Carcinogenesis. 2010 April; 31(4):552-558.). The pro-survival activity in sarcoma can—at least in part—be ascribed to the role of DYRK1B in promoting the inactivation of reactive oxygen species (ROS). DYRK1B is able to increase the expression of ROS detoxifying enzymes including superoxide dismutases 2 and 3 (Deng et al., Cancer Res. 2009 Apr. 15; 69(8):3317-3324; Hu and Friedman, Genes Cancer. 2010 Aug. 1; 1(8):803-811.). This may also explain the enhanced sensitivity of DYRK1B-depleted cancer cells to certain chemotherapeutic drugs such as cisplatin known to increase toxic ROS levels (Hu and Friedman, Genes Cancer. 2010 Aug. 1; 1(8):803-811.).
Hedgehog Signaling in Cancer Therapy
The Hedgehog (HH)/GLI signal transduction pathway is a key regulator of multiple developmental processes. Uncontrolled activation of HH/GLI signaling is a common feature of many human malignancies including cancers of the brain, skin, gastro-intestinal tract, prostate, breast lung, muscle and bone (reviewed in (Beachy et al., Nature. 2004 Sep. 23; 431(7007):402; Beachy et al., Nature. 2004 Nov. 18; 432(7015):324-331; Epstein, Nat Rev Cancer. 2008 October; 8(10):743-754; Kasper et al., J Clin Invest. 2012 Feb. 1; 122(2):455-463; Kasper et al., Eur J Cancer. 2006 March; 42(4):437-445; Merchant and Matsui, Clin Cancer Res. 2010 Jun. 15; 16(12):3130-3140; Ng and Curran, Nat Rev Cancer. 2011 May 26; 11(7):493-501; Ruiz i Altaba et al., Trends Cell Biol. 2007 September; 17(9):438-447; Ruiz i Altaba et al., Nat Rev Cancer. 2002 May; 2(5):361-372; Scales and de Sauvage, Trends Pharmacol Sci. 2009 June; 30(6):303-312; Teglund and Toftgard, Biochim Biophys Acta. 2010 April; 1805(2):181-208; Theunissen and de Sauvage, Cancer Res. 2009 Aug. 1; 69(15):6007-6010.).
Precise reversible regulation of Hedgehog signaling is a complex process and mandatory for proper normal development of both invertebrate and vertebrate organisms (for detailed reviews see (Huangfu and Anderson, Development. 2006 January; 133(1):3-14; Ingham and McMahon, Genes Dev. 2001 Dec. 1; 15(23):3059-3087; Teglund and Toftgard, Biochim Biophys Acta. 2010 April; 1805(2): 181-208.)).
In the absence of HH ligand, HH signaling is repressed by the activity of the HH receptor Patched (PTCH), a twelve-transmembrane domain protein whose intracellular localization is concentrated at the base of the primary cilium, a single antenna-like cell surface compartment that coordinates HH signal transduction. Unliganded PTCH prevents the translocation of the G-protein coupled receptor-like protein and essential pathway effector Smoothened into the primary cilium (Corbit et al., Nature. 2005 Oct. 13; 437(7061):1018-1021; Rohatgi et al., Science. 2007 Jul. 20; 317(5836):372-376; Rohatgi and Scott, Nat Cell Biol. 2007 September; 9(9): 1005-1009.). This leads to protcolytic cleavage of the latent zinc finger transcription factors GLI3- and to some extent also of GLI2- into C-terminally truncated repressor forms (GLIR). GLIR formation involves preceding and sequential phosphorylation by protein kinase A (PKA), glycogen synthase kinase 3-beta (GSK) and casein kinase I (CKI) (Price and Kalderon, Cell. 2002 Mar. 22; 108(6):823-835.) as well as a functional primary cilium (Smith and Rohatgi, Sci Signal. 2011 Jan. 11; 4(155):mr1; Wang et al., Cell. 2000 Feb. 18; 100(4):423-434; Wen et al., Mol Cell Biol. 2010 April; 30(8):1910-1922; Wong et al., Nat Med. 2009 September; 15(9):1055-1061.). Following processing, GLIR translocates to the nucleus to bind to HH target gene promoters and repress target gene expression (Aza-Blanc and Kornberg, Trends Genet. 1999 November; 15(11):458-462; Aza-Blanc et al., Cell. 1997 Jun. 27; 89(7):1043-1053.). GLI signals are also negatively regulated by proteasome-mediated degradation of GLI and by binding to Suppressor of Fused (SUFU), which sequesters GLI proteins in the cytoplasm and also contributes to GLI processing in the primary cilium (Humke et al., Genes Dev. 2010 Apr. 1; 24(7):670-682; Kogerman et al., Nat Cell Biol. 1999 September; 1(5):312-319.).
The therapeutic relevance of targeting HH/GLI signaling in human cancers with genetic, ligand-independent activation of HH/GLI signaling has recently been demonstrated for basal-cell carcinoma (BCC) and medulloblastoma. In both malignant entities, inhibition of the essential HH pathway effector Smoothened had a dramatic therapeutic benefit (Rudin et al., N Engl J Med. 2009 Sep. 17; 361(12):1173-1178; Skvara et al., J Invest Dermatol. 2011 August; 131(8):1735-1744; Von Hoff et al., N Engl J Med. 2009 Sep. 17; 361(12):1164-1172.). Whether Smoothened antagonists will display therapeutic efficacy in HH ligand dependent cancers remains to be shown. Ongoing clinical trials with Smoothened antagonists from different pharmaceutical companies will eventually answer the question of the clinical efficacy of targeting Smoothened in Hedgehog associated malignancies (Aberger et al., Vitam Horm. 2012; 88:25-54; Lin and Matsui, Onco Targets Ther. 2012; 5:47-58; Ng and Curran, Nat Rev Cancer. 2011 May 26; 11(7):493-501; Scales and de Sauvage, Trends Pharmacol Sci. 2009 June; 30(6):303-312.). Clinical studies with small molecule Smoothened inhibitors to treat patients with metastatic colorectal cancer, ovarian cancer or pancreatic cancer failed to demonstrate therapeutic efficacy of Smoothened antagonists in combination with currents treatment regimens (Ng and Curran, Nat Rev Cancer. 2011 May 26; 11(7):493-501.).
One of the reasons for the lack of therapeutic efficacy of Smoothened inhibitors may be explained by Smoothened-independent activation of GLI transcription factors in different cancer entities such as pancreatic cancer, melanoma or Ewing's sarcoma. This non-canonical activation of GLI transcription factors can be induced by a variety of signals frequently hyperactive in malignant cells including TGF-b/SMAD, RAS-MEK/ERK, PI3K/AKT, EGFR signaling or the EWS-FLI1 oncogene (reviewed in (Aberger et al., Vitam Horm. 2012; 88:25-54; Mangelberger et al., Front Biosci. 2012 Jan. 1; 17:90-99; Stecca and Ruiz, J Mol Cell Biol. 2010 April; 2(2):84-95.).
Regulation of HH/GLI Signaling by DYRK Family Members
The first regulatory interactions between DYRK family members and the HH/GLI pathway came from studies of DYRK1A and its impact on the transcriptional activity of the GLI zinc finger transcription factors mediating the transcriptional output of HH pathway activation. Using reporter gene based assays, Mao et al. have shown that DYRK1A is able to enhance the activity of the GLI1 activator and stimulate HH target gene expression, respectively. DYRK1A can phosphorylate GLI1 in vitro and enhance the nuclear level of GLI1. Direct modification of GLI1 and enhanced nuclear localization in response to DYRK1A activity are likely to account for the enhanced expression of HH target genes (Mao et al., J Biol Chem. 2002 Sep. 20; 277(38):35156-35161.).
While DYRK1A enhances GLI activity, the class II DYRK family member DYRK2 acts as negative regulator of GLI activity. DYRK2 can directly phosphorylate GLI2 and GLI3 resulting in destabilization of GLI2/3 and enhanced proteasome-dependent degradation. Mutation of the DYRK2 substrate phosphorylation sites S384 and S1011 in GLI2 rendered GLI2 resistant to DYRK2 mediated inhibition of transcriptional activity and proteasomal degradation (Varjosalo et al., Cell. 2008 May 2; 133(3):537-548.).
Analysis of DYRK1B function in HH-unresponsive RAS mutant pancreatic cancer cells revealed another regulatory mechanism by which DYRK kinases can affect the activity of HH signaling. Lauth et al. (Nat Struct Mol Biol. 2010 June; 17(6):718-725) provide evidence that DYRK1B is involved in an autocrine-to-paracrine shift of HH signaling triggered by mutant RAS. This study suggests that oncogenic RAS signaling in pancreatic cancer cells increases HH ligand expression though at the same time it also prevents autocrine HH pathway activation (Lauth et al., Nat Struct Mol Biol. 2010 June; 17(6):718-725.). RAS signaling therefore contributes to paracrine HH signaling, with tumor cells representing the signal source and adjacent stroma cells the signal-receiving compartment (Yauch et al., Nature. 2008 Sep. 18; 455(7211):406-410.). Like RAS, expression of the RAS effector DYRK1B in HH activated mouse fibroblasts inhibited HH signaling, suggesting that DYRK1B can act downstream of RAS to prevent autocrine HH signaling. Further, RNAi knockdown of RAS and DYRK1B in RAS mutant pancreatic cancer cells both led to a GLI2-dependent increase in GLI1 mRNA expression (Lauth et al., Nat Struct Mol Biol. 2010 June; 17(6):718-725.). The detailed mechanisms of HH pathway inhibition by DYRK1B remain unknown.
Together, these reports show that DYRK2 and DYRK1B can have a repressive effect on HH/GLI signaling while DYRK1A functions as positive regulator of GLI transcriptional activity.